idea for Neo 2nd generation: Accelerometer

Tim Newsom cephdon at
Fri Jan 26 18:41:50 CET 2007

On Fri, 26 Jan 2007 9:12, Steven Milburn wrote:
> yes,  accelerometers measure acceleration.  The first derivative of 
> acceleration is velocity.  Granted errors in the accelerometer compound 
> when deriving velocity, but you've usually got GPS information to 
> calibrate against (As Jeff was saying).  A typical dead reckoning 
> system always knows your current velocity,  and when GPS goes away, it 
> can apply changes to the velocity by knowing only your current 
> acceleration.  The errors associated with this would typically be small 
> enough to not matter during the length of a tunnel or building related 
> GPS blackout.
> So, an accelerometer (actually, three) is all that is needed for a 
> complete navigation with dead reckoning.
> --Steve

Ok Steve.  I grant you that the first derivative of acceleration is 
velocity... How do you propose to gain any velocity information when the 
acceleration measured is zero as would be the case if you are at a 
constant velocity?  This is why I am saying you would need some better 
source for velocity.
  I grant you that the times when a car is at a constant velocity may be 
few... Or it may be that when on cruise control on any flat road you may 
actually see zero or almost zero acceleration.

OTOH, detecting the direction that a person turned with the 
accelerometer may be very useful in dead reckoning.



More information about the community mailing list